In a world where fake news spreads at breakneck speed, it’s getting harder and harder to separate fact from fiction. And despite countless hours of work and millions of dollars invested in projects and policies aimed at promoting truthful journalism, the situation remains the same. far from ideal—and that was before artificial intelligence could take part in creating fake images that looked real.
in fact, According to the European Data Protection Supervisor, fake news is considered free speech and it is almost impossible to combat misinformation because “manually dealing with the mass of fake news spread on social media is not possible”. is.
Fortunately, artificial intelligence can also help uncover fake news. This power comes primarily from the explosive growth of large language models like GPT-4.
For example, an anonymous developer launched AI Fact Checker, a tool created to fact-check information using AI. When a user enters a claim into the checker, the platform searches reliable sources on the internet, analyzes the data and compares the information with the claims provided. Then, determine whether the claim or fact is true or false, and provide the sources that support your conclusion.
Decryption tested the tool’s capabilities and demonstrated a 100% accuracy level in fact-checking recent news, historical events, and other information. But when it came to consulting news on the prices of goods, services and investment vehicles, the tool stumbled and began to confuse predicted news with actual price movements.
Other AI tools target the same problem. Google’s Fact Check Tools, Full Fact, and FactInsect are among the most popular.There are even decentralized alternatives like fact protocol And the now-deceased Sybil. But Microsoft’s AI-powered browser and search engine stand out for their ease of use and accuracy.
The software giant integrated GPT-4 into the Edge browser, making AI bots readily available to users. Unlike ChatGPT (which cannot search the web for information and has a fixed data set before 2021), the new Bing with GPT-4 can browse the web so you can get accurate, up-to-date answers instantly I can provide it and provide a link. Be the authoritative source for any questions, including questionable fact-checks or rants.
To verify fake news using GPT-4 with Bing, download the latest version of the Edge browser and click the Bing logo icon in the upper right corner. A sidebar menu opens with three options. Then select the chat option, ask what you want to know, and Microsoft’s AI will answer.
Microsoft’s GPT-powered AI validates the Decrypt headline. We do not spread fake news.
the user is[Insights]You can also click Options. Microsoft’s GPT-powered AI provides relevant information about websites publishing news. This includes a description of the topic covered, related themes, page his traffic, and general criticism. Unlike the AI Fact Checker Tool, it doesn’t give you specific “this is right” or “this is wrong” answers, but it does give you enough information to reach a conclusion.
back side of coin
AI can track and compare multiple sources in seconds, but using AI algorithms to verify news also comes with risks. Drawbacks of AI include:
Training a flawed model: If an AI algorithm is trained on inaccurate or biased data, its performance can be adversely affected and lead to erroneous results. It’s important to ensure that the data used to train your AI algorithms is accurate and representative.
AI Hallucinations: AI algorithms can generate information that looks plausible but has no real basis. These hallucinations can lead to false conclusions when reviewing the news.
Manipulation Vulnerability: AI algorithms can be susceptible to attacks and manipulation, such as injecting fake or biased data into the learning process. These are somewhat equivalent to a 51% attack, but in the context of AI. More wrong reinforcements are used, causing the model to assume wrong data to be correct.
This is of particular concern for models that rely on human interaction or are subject to a central entity managing them…wThis is exactly what happens with OpenAI and ChatGPT.
